During 20-21 August 2024, the Northeastern Indian state of Tripura, which is mountainous, received heavy rainfall leading to destructive flooding and loss of life. Downstream of this region is the low-lying country of Bangladesh formed by the deltaic deposits of the Ganges and Brahmaputra rivers. The lack of preparation for such a rapid onset of extreme flooding impacted more than a million people forcing 80,000 individuals into the 557 makeshift relief camps set up by the Tripura State Government (Reuters, August 24 2024).
As the flood travelled downstream crossing the international boundary with Bangladesh, the flood quickly became a transboundary flood, catching the inhabitants of Bangladesh also off guard due to lack of early warning.
As news of Tripura having to open the gate of its sole hydroelectric Dumboor dam became public, a flood of disinformation ensued within Bangladesh as the actual flooding wreaked havoc. Some of this disinformation presented itself as conspiracy theories or conjectures, such as: Did India intentionally flood Bangladesh by opening the gate of Dumboor dam?”, “Was it really such a heavy and widespread rainfall event that the dam played no role?” and “Could India (Tripura Government) not have warned downstream authorities in Bangladesh?”
Tackling disinformation with satellites
Amid such wild speculation shared widely by the public, as the catastrophic flooding needed a rapid disaster response and relief in low-lying Bangladesh, we asked ourselves, on 22 August 2024, “what do satellites in space tell us about this evolving transboundary flooding given that satellites can see both sides of the border without bias?” (Hossain and Katiyar, 2006).
We explored the use of satellites to understand the evolving transboundary flooding situation on the ground, the potential role of upstream dam and the skill of weather forecast. The goal of this exploration was to eventually develop a tool that can improve livelihoods through sustainable application of earth science, remote sensing and advanced information technology. This goal happens to be a core research philosophy of the group we are affiliated with at the University of Washington, known as SASWE (https://www.saswe.net).
In other words, given the plethora of big data and advanced information technology with improved modeling, we asked: “what can Earth observations and data science do to make a positive impact for everyone on both sides of the border?”
This question takes particular importance for two reasons. The first reason is personal wherein the flood impacted the hometown of the first author (Pritam Das) who is from Agartala, Tripura. The second reason is the Surface Water and Ocean Topography (SWOT) satellite mission that was launched in December 2022 started streaming data for the public on water on the surface.
SWOT makes the Tripura-Bangladesh situation particularly unique and exciting as it can measure, for the first time in satellite remote sensing history, water elevation and inundation extent to provide a more powerful and all-weather synopsis of the evolving flood situation and reservoir state.
What did Earth Observation data from satellites and modelling reveal?
Figure 3 shows if the widely available precipitation forecasts from global weather prediction models, such as the Global Ensemble Forecasting System (GEFS) produced by NOAA, revealed anything insightful. It also shows where most of the precipitation fell based on observation from the Global Precipitation Measurement (GPM) mission data product called IMERG.
We are able to make the following observations to dispel the flood of disinformation:
- Most of the precipitation and runoff occurred downstream and outside of Dumboor dam catchment. This can be inferred by comparing the left panel’s lower right image with the location of the Dumboor dam that is located above the core of where precipitation took place.
- Precipitation forecast from widely available global numerical weather prediction (NWP) models such as GEFS from NOAA missed the intensity of the precipitation at 3 day lead time. We can infer this by comparing the lower right sub-plot of left panel with the right panel plots and the red arrow.
- Global NWP models can suffer from poor spatial precision of the most intense precipitation events.
It is clear from the data that since most of the precipitation fell downstream of the dam in India, the dam itself could not have the strongest impact on exacerbating the flooding downstream in India.
Next, we rapidly prototyped the satellite based reservoir tracking system that can track and forecast reservoir state and reservoir release along with broader inundation state using SWOT inundation map. Figure 4 shows the pre-flood water level conditions according to SWOT on 16 August 2024. SWOT inundation data reveals that the water level at the border region between India and Bangladesh were higher than the surrounding flat regions, particularly in the southeastern regions of Bangladesh (locally known as Feni that was most affected by this flooding). This higher water level of rivers at the border region may have been caused by the preceding monsoon rains, thereby leading to conditions ripe for the transboundary flooding during Aug 21-22 to become catastrophic.
The satellite-based reservoir tracking and forecasting system is called Reservoir Assessment Tool (RAT) Tripura and is currently operationalized at https://depts.washington.edu/saswe/tripura. RAT Tripura uses the concepts of open source modelling, cloud computing and a reservoir tracking concept formalised and made publicly available as scalable software by Minocha et al. (2024).
RAT-Tripura is a modeling system that now runs in near real-time using lowest latency satellite data and produces routine updates of the flooding and reservoir situation (inflow and outflow) to empower water management agencies on both sides of the India-Bangladesh border.
Figure 5 shows that RAT-Tripura, if it was already in place as a practitioner tool for flood management agencies of India and Bangladesh, and if the focus was on trends of the transboundary river flow with an eye on the upstream Dumboor dam, livelihoods could have been impacted with better early warning up to 2-3 days.
More cooperation, less disinformation
RAT-Tripura (Reservoir Assessment Tool)-Tripura is now actively under consideration and investigation as an operational decision-assisting tool by the Bangladesh Water Development Board’s (BWDB) Flood Forecasting and Warning Center (FFWC). As the flood was evolving, the authors had the opportunity to reach out to the staff of BWDB Director General and FFWC on 25 August 2024, and share the tool. Around the same time, the Ministry of Water Resources and Environment of Bangladesh under the current interim government led by Professor Yunus, also had a chance to learn about the SWOT satellite emission, RAT-Tripura and what satellite-based Earth observations revealed during the peak of the flood event. The counterpart agencies in the Indian state of Tripura were also reached out to.
Currently, conversations continue with BWDB and FFWC on how to make the most optimal use of such satellite data and open-source reservoir tracking tools through capacity building of their staff.
As the call for more transboundary cooperation and treaties for sharing information and water ramp up, we believe our efforts and tools will play a vital role in understanding how satellites from space can provide an unbiased view of the ground situation to fight disinformation and promote transboundary collaboration.
Pritam Das, Sanchit Minocha, Shahzaib Khan and Faisal Hossain, University of Washington, Seattle, WA 98195, US.
